jbmusso / awesome-graph Goto Github PK

Hey nice resources collected. Thanks!
I truly believe you should have a look at TerminusDB,

https://github.com/terminusdb/terminusdb

Hey, I found this graph database: https://github.com/dgraph-io/dgraph

Memgraph cloud is fully managed on AWS. Can it be added to this list?

Created Dydra entry with short explanation.

Nebula is A new opensource graph database, Our opensource repositry is here: https://github.com/vesoft-inc/nebula.

The group of graph databases that use datalog variants as query languages grows everyday - in fact, I would suggest that it is swiftly becoming the dominant query approach in the new generation of graphs dbs. These new superstars all use datalog:

TerminusDB
Grakn
Crux

Our lab is developing the PyGSP, a python package for graph signal processing. It may be relevant to your graph computing list. Our approach is different from the other packages as we take a signal processing point-of-view to computations. The Emerging Field of Signal Processing on Graphs provides a nice overview.

What do you think of adding Graph Convolutional Networks in a new section on ML and deep learning?

Here are some references:

• https://tkipf.github.io/graph-convolutional-networks/
• https://github.com/tkipf/gcn (implementation of the above)
• https://towardsdatascience.com/how-to-do-deep-learning-on-graphs-with-graph-convolutional-networks-7d2250723780

Could be interesting not to just mention the stores, but also framweorks for visualization / app development, like (the list below is probably not complete! )

• Graphileon ( that's us)
• Structr
• Linkurious
• Keylines

Was browsing and wondered if Anzograph - an OLAP graph database should be on your list?

https://www.cambridgesemantics.com/product/anzograph/

It points to http://tinkerpop.apache.org/docs/currentg/#neo4j-gremlin which returns a 404.

We are looking to use a graph DB. Our data set has about 500M vertices and 2Bn directed edges. Each edge has an index from 0 to 1M (i.e. on average 2000 edges share the same index). We can have parallel edges with different indices. We need to be able to query for a desired index range. Also, each edge has another value attribute, and each vertex has one scalar attribute. We would have to add more vertices and edges as time goes on. We would like to host this on a single server with large memory.

• Which graph databases would scale for this data set?
• How would you model such a data set in a graph DB? Should it be a one giant graph or maybe a series of 1M sub-graphs (e.g. one graph per each index value)?